Hospitals across the country are investing heavily in artificial intelligence tools, lured by vendor promises of rapid financial returns, streamlined workflows and newfound clinical capacity. But according to several chief medical information officers, the reality on the ground is far more nuanced.
When Becker’s asked CMIOs where they see the biggest gap between how AI’s return on investment is marketed and what they’ve experienced, a consistent theme emerged: The real ROI is slower, more complex and often rooted in human factors like clinician well-being — not just dollars saved.
David Whitling, MD, CMIO at Boulder (Colo.) Community Health, said early marketing set unrealistic expectations for clinical time savings.
“This technology undoubtedly saves time, but our data shows a modest improvement only, likely because clinicians wisely use that time for other patient care tasks,” he told Becker’s. “Health systems that adopted this technology by relying on the formula ‘AI means time saved, which means opportunity to see more patients,’ have likely been disappointed.”
Many leaders noted that while AI tools — particularly ambient listening and documentation assistants — are proving valuable, they don’t deliver the quick, dramatic returns often promised by vendors. Usman Akhtar, MD, CMIO of Virginia Hospital Center in Arlington, said vendors often promote immediate, tangible ROI, when in reality the benefits are slower to be realized, less direct and more human-focused: reducing burnout, improving provider satisfaction, enhancing note quality and gaining incremental improvements in coding capture.
“These are still very real,” he said. “But they rarely align with the ‘30% cost reduction in 6 months’ narrative.”
Clay Callison Jr., MD, CMIO at Knoxville-based University of Tennessee Medical Center, echoed that point, noting that ambient listening may enhance patient access but its most substantial benefit lies in improving clinician quality of life and reducing burnout — results that take more than a few months to emerge. Matthew Anderson, MD, CMIO of Scottsdale, Ariz.-based HonorHealth, added that from a clinical productivity perspective, the incremental benefits of AI tools simply need a longer timeline to show meaningful impact.
For many CMIOs, the most significant returns they have observed are not financial at all. Instead, they center on cognitive relief and job satisfaction.
“We are not seeing the transformative gains in clinical efficiency that were promised,” said Joseph Evans, MD, chief health information officer at Norfolk, Va.-based Sentara Health. “However, what we are seeing is a profound and invaluable return in the form of improved clinician well-being. The reduction in cognitive load and documentation burden is directly translating to lower burnout rates and higher professional satisfaction. This ‘well-being ROI’ is critically important, even if it’s harder to quantify on a balance sheet.”
That sense of relief is echoed by others. Kristin Conley, DO, CMIO at Frederick Health in Maryland, described ambient documentation as a “hip-hop-powered sidekick” that brings joy back to the workday, likening the experience to 90s classics: “‘Joy and Pain’ — with ambient there’s more joy in your day and less pain in your workflow.”
Still, measuring these qualitative benefits remains difficult. “We get a lot of great feedback on AI in terms of getting home on time or reducing cognitive load, but it’s pretty challenging to directly map that to a line item on a budget,” said Neal Chawla, MD, CMIO at Raleigh, N.C.-based WakeMed.
Implementation challenges add another layer of complexity. Several leaders noted that AI is rarely “plug and play.” Matthew Shafiroff, MD, CMIO at White Plains Hospital in New York, said real ROI happens when organizations communicate the “why,” simplify workflows for clinicians and staff, and iterate with them to get it right.
“Skip the change management piece, and even great AI will sit unused,” he added.
Amer Saati, MD, CMIO at Roseville, Calif.-based Adventist Health, said the organizations seeing success are those investing as much in workflow redesign and change management as in the technology itself. Elie Razzouk, CMIO at Altamonte Springs, Fla.-based AdventHealth, added that vendors often highlight performance under idealized conditions, which rarely reflect real-world outcomes. The true value of AI, he said, lies in its ability to produce meaningful, sustainable impact in actual clinical settings.
Some CMIOs suggested that the problem lies not just in timelines or implementation — but in how ROI is defined.
“For clinicians not using AI going forward, it would be equivalent to using carbon-copy order forms today: inefficient and unsafe,” Nadeem Ahmed, MD, CMIO at The Valley Health System in Paramus, N.J., said. “AI has not been marketed in this manner, and maybe it should be.”
Mark Mabus, CMIO at Fort Wayne, Ind.-based Parkview Health, cautioned against overgeneralizing from early pilots.
“Many of the early ROI claims for GenAI came from carefully selected pilot groups that don’t represent the broader provider population, making those benchmarks unrealistic for real-world adoption,” he said.
